Hospital Type and Patient Outcomes: An Empirical Examination Using AMI Re-Admission and Mortality Records

Abstract

This paper investigates whether there are differences in patient outcomes across different types of hospitals using patient-level data on re-admission and mortality associated with acute myocardial infarction (AMI). Hospitals are grouped according to their ownership status (private, teaching, non-teaching) and their location (metropolitan, country and remote country). Using data collected from 130 Victorian hospitals on 19,000 patients admitted to a hospital with their first AMI between January 2001 and December 2003, we consider how treatment affects the likelihood of various outcomes based on unplanned re-admission and mortality. A hazard rate model is used to assess the effect of hospital type on patient outcome. Control variables included in the estimating model are patient-level characteristics such as age, gender, co-morbidity, country of birth, and indigenous, marital and socio-economic status. We find that there are significant differences across hospital types in the outcomes observed for patients presenting with their first AMI - private hospitals persistently outperform teaching, non-teaching and country hospitals. Interestingly, we find that result is that the impact of hospital type is quite robust to the definition of patient outcomes that we adopt and our attribution strategy, but not to whether we include multiple-hospital patients.

Keywords: mortality, acute myocardial infarction, hospital performance, hazard model

JEL Classification: C41, I12